Perception-Distortion Trade-Off in the SR Space Spanned by Flow Models
Authors: Cansu Korkmaz, A. Murat Tekalp, Zafer Dogan, Erkut Erdem, Aykut Erdem
Venue: IEEE International Conference on Image Processing (ICIP) 2022, Bordeaux, France, 16-19 October 2022
DOI: 10.1109/ICIP46576.2022.9897761
Overview
This work explores the perception-distortion trade-off in super-resolution using flow models. The paper provides insights into balancing perceptual quality and distortion metrics within the solution space spanned by flow-based generative models, offering a comprehensive analysis of the trade-offs involved in super-resolution tasks.
Key Contributions
- Analysis of perception-distortion trade-off in flow model space
- Comprehensive evaluation framework
- Insights into balancing perceptual and distortion metrics